Gui Workflow Editor User Guide

Getting there

You can either:

• Edit an existing campaign
  • Go to the Campaigns/Tagged Campaigns page from the navigation bar/menu.
  • Pick the [Edit Campaign] button for one of the tagged campaigns
• Clone an existing campaign
  • Go to the Campaigns/Tagged Campaigns page from the navigation bar/menu.
  • Fill in from/to replacement strings (i.e MC9 -> MC10) for one of the tagged campaigns
  • Click the [Clone Campaign] button
• Clone a template/sample Campaign
  • Go to the Template Campaigns link in the navigation bar/menu (NOTE: not yet implemented)
  • fill in a replacement string for the "generic" in the campaign
  • Click the [Clone Campaign] button

You should now be at a screen that looks like:

You have

• a "Default Values" box with values common accross your Campaign Stages.
  • you can click the [icon] box to see the form to edit the values,
  • and then click the [icon] box to hide the form again
• a diagram of the campaign stages with dependencies connecting them.
  • each campaign stage has an [grid icon] you click to toggle the view of the propreties edit form
  • each dependency link has an [grid icon] you click toggle the view of its properties edit form
• Above the diagram are buttons to
  • create a new stage in the campaign
  • link two stages with a dependency
• Notes on stages
  • stages are draggable so you can make room for new stages.
  • the edit form also has an [recyle icon] you click to delete that item
• There are also boxes for the campaign's
  • job type(s) and
  • login/launch templates

Sample task tutorials

Adding a new campaign stage in the middle

• Go into the editor
• drag the campaign stages following where you want to add it over to make room
• remove the existing dependency
  • click the [grid icon] on the dependency line
  • click the [recylce icon] on that form
  • confirm the popup
• click "New Stage"
• Enter its name into the popup
• Click the previous stage to highlight it
• Click the new stage to highlight it
• Click on "Add dependency"
• Click the previous stage again to un-highlight it
• Click the following stage to highlight it
• Click the "Add dependency"
• Click Save

Changing the software version in the whole campaign

• Go into the editor
• click on the Default Values [grid icon]
• change the software version
• click the [grid icon] again to close the form
• Click Save
• you can click the individual stage [grid icon] to open/close the form to make sure they havent overridden it

Changing Data Management parameters

A POMS campaign can be configured to use either SAM or Data Dispatcher data management systems; you can also use the SAM option for campaigns that don't really use a data management system. You can select which one by

• double clicking on the oval Campaign bubble,
• clicking the [Edit] button on the "data handling service" entry, and
• choosing either SAM or Data Dispatcher in the pulldown, followed by
• clicking Accept and then
• clicking [Save] for the overall campaign.
  When the save completes, you can fill in default data management parameters in the oval Campaign, or in specific campaign stages, by double clicking on the appropriate icon, as below.

Changing SAM parameters

For campaigns with SAM, you when you click on [Edit] on the "sam settings" entry, you only have one parameter to set, the "SAM dataset or split data", which is usually the name of a SAM dataset the campaign stage should process; If you are using a dataset split type, this dataset name is the dataset to be split,(although some split types actually take other data, such as a SAM project id, etc. which you would also put in here.) There are not otherwise many parameters to fill in here, because the SAM setup expects the job launch scripts to actually start SAM projects, and POMS does not do it for them.

Changing Data Dispatcher parameters

For campaigns using Data Dispatcher, when you click on [Edit] in the "data dispatcher settings" line, you get a popup with the following parameters you can change:

• Project ID existing project from which to get files -- use either this or one or more of the next 4 items.
• Dataset Query query for file list to start a new project when launching jobs
• Idle Timeout idle timeout for new projects
• Worker Timeout worker timeout for new projects
• Virual Mark project as "virtual" (ignores metacat replica status) for new projects
• Load Limit Passed to fife_wrap for "--data_dispatcher" file processing loop -- try to load this many files at once and hand them to one executable invocation
• Stage Methodology Recovery behavior -- can currently be:
  • Regular -- new project per job launch
  • OneP -- Use the same data_dispatcher project to recover files not marked permanently failed
• Recovery Mode can be either
  • Standard
  • Aggressive